Microsoft's Top-secret AI for US Spies
Plus: Google DeepMind releases AlphaFold 3, OpenAI’s shares how it teaches it’s models to behave with Model Spec.
Hello Engineering Leaders and AI Enthusiasts!
Welcome to the 271st edition of The AI Edge newsletter. This edition brings you Microsoft’s secretive AI it built for US spies.
And a huge shoutout to our amazing readers. We appreciate you😊
In today’s edition:
🕵️♀️ Microsoft developed a secretive AI service for US spies
🚀Google DeepMind and Isomorphic Labs introduce AlphaFold 3🧠OpenAI’s Model Spec shares how it teaches it’s models to behave
📚 Knowledge Nugget: The mobile S-curve ends, and the AI S-curve begins by
Let’s go!
Microsoft developed a secretive AI service for US spies
Microsoft has developed a top-secret generative AI model entirely disconnected from the internet so US intelligence agencies can safely harness the powerful technology to analyze top-secret info. The model based on GPT-4 is now live, answering questions, and will also write code.
Microsoft spent 18 months developing the model, which is "air-gapped" to ensure it is secure. This is the first time a model is fully isolated– meaning it's not connected to the internet but is on a special network that's only accessible by the U.S. government.
It can read and analyze files but cannot learn from them to stop sensitive information from entering the platform. It is yet to be tested and accredited by the intelligence agencies.
Why does this matter?
Intelligence agencies all over the world have been racing to be the first to harness generative AI. I guess we know who’s going to be the winner. If this AI tool is successful, it will fundamentally change the way intelligence agencies operate.
Google DeepMind and Isomorphic Labs introduce AlphaFold 3
AlphaFold 3 is a revolutionary model that can predict the structure and interactions of all life’s molecules with unprecedented accuracy.
For the interactions of proteins with other molecule types, it sees at least a 50% improvement compared with existing prediction methods, and for some important categories of interaction it has doubled prediction accuracy. AlphaFold 3’s capabilities come from its next-generation architecture and training that now covers all of life’s molecules.
Google DeepMind has also newly launched AlphaFold Server. It is a free platform that scientists worldwide can use for non-commercial research. With just a few clicks, biologists can harness the power of AlphaFold 3 to model structures composed of proteins, DNA, RNA and a selection of ligands, ions and chemical modifications.
Why does this matter?
Previous AlphaFold models have been used to make discoveries in many areas. But AlphaFold 3 takes us beyond proteins to a broad spectrum of biomolecules. This leap could unlock more transformative science, from developing bio-renewable materials and more resilient crops, to accelerating drug design and genomics research.
OpenAI’s Model Spec shares how it teaches it’s models to behave
OpenAI has shared the first draft of Model Spec, a new document that specifies its approach to shaping desired model behavior and how it evaluates tradeoffs when conflicts arise. It brings together documentation used at OpenAI today, their experience and ongoing research in designing model behavior, and more recent work, including inputs from domain experts, that guides the development of future models.
OpenAI intends to use the Model Spec as guidelines for researchers and AI trainers who work on reinforcement learning from human feedback (RLHF). It will also explore to what degree its models can learn directly from the Model Spec.
Why does this matter?
All AI makers are looking for efficient methods to control the behavior of their models, wanting them to follow guidelines but not reject normal requests. OpenAI offers a window into the often opaque process of shaping AI behavior.
While it isn’t showing its whole hand here, Model Spec can be helpful to users and developers to see the high-level rules that indirectly govern ChatGPT and other models.
Enjoying the daily updates?
Refer your pals to subscribe to our daily newsletter and get exclusive access to 400+ game-changing AI tools.
When you use the referral link above or the “Share” button on any post, you'll get the credit for any new subscribers. All you need to do is send the link via text or email or share it on social media with friends.
Knowledge Nugget: The mobile S-curve ends, and the AI S-curve begins
In an interesting post
discusses the contrast between the mobile app ecosystem, which is nearing the end of its growth curve or "S-curve", and the emerging generative AI ecosystem, which is just beginning its own S-curve of rapid growth and innovation. He also argues that:Early in an S-curve, products just need to "work" to attract users due to novelty, but late in the curve, products must be radically different to succeed as user expectations are higher.
Early S-curve products benefit from novelty effects driving rapid user growth but high churn, while late S-curve requires focusing on retention over acquisition.
Investors fund early S-curve opportunities for their growth potential, while late S-curve startups need innovative design and go-to-market strategies to compete.
The rapid cadence of new technology adoption means S-curves rise and fall faster, requiring agility from founders to succeed across different curve stages.
Why does this matter?
In essence, it contrasts the dynamics and strategies needed to win in an emerging vs mature technology ecosystem. This provides both opportunities and challenges that startups and investors in the AI space need to be aware of and strategize around.
What Else Is Happening❗
📝Copilot for Microsoft 365 to get auto-complete and rewrite to improve prompts
In coming months, Microsoft Copilot will be updated with new features like auto-complete and ‘elaborate your prompt’ that offer suggestions to improve AI prompts. It aims to solve the problem of coming up with good prompts for generative AI. (Link)
🏢New AI data center to be built at the failed Foxconn project site in Wisconsin
President Joe Biden announced an AI data center to be built on the same site as the failed Foxconn project in Racine, Wisconsin. According to a White House press release, Microsoft is investing $3.3B in the project, creating up to 2,000 permanent jobs. (Link)
🤔Sam Altman says we are not taking AI’s impact on the economy seriously
At a Brooking's Institute panel about AI and geopolitics on Tuesday, Altman said the discussions around AI's effect on the economy– like how it may lead to mass job replacement– died down this year compared to last. He said if we don’t take these concerns seriously enough going forward, it could be a massive issue. (Link)
✒️Typeface Arc replaces prompts; uses AI agent approach to ease marketing workflows
It is launching Typeface Arc technology, which enables a user to state a high-level marketing objective and then have the system automatically plan and generate all the assets, including emails, images, and notifications that are all connected. (Link)
🎮Altera’s gaming AI agents get backed by Eric Schmidt, Former Google CEO
Altera is the newest startup joining the fray to build a new guard of AI agents. It raised $9 million in an oversubscribed seed round, co-led by Eric Schmidt’s deep-tech fund, First Spark Ventures and Patron, the seed-stage fund co-founded by Riot Games alums. (Link)
New to the newsletter?
The AI Edge keeps engineering leaders & AI enthusiasts like you on the cutting edge of AI. From machine learning to ChatGPT to generative AI and large language models, we break down the latest AI developments and how you can apply them in your work.
Thanks for reading, and see you tomorrow. 😊